修复 Bittensor (TAO) 子网利益不一致问题的 10 项提案

Covenant decided to leave the TAO ecosystem, taking 8 figs in capital with them The biggest friction point to adoption here is the misalignment in subnet and network interest

Very few institutions will deploy serious capital until a tangible solution is in place

Here are 10 suggestions that could alleviate this problem:

• Lock-based subnet ownership as const suggests

Founders lock their tokens over multi-year schedules visible on chain. If you believe in what you're building this shouldn't be a problem. Investors get advance warning of any unlock and can reprice accordingly

• Protocol-level IP retention. If a model is trained using Bittensor emissions and network compute the weights and outputs should be retained by the network

Covenant trained a 72B param model on Bittensor resources and took it with them. That shouldn't be possible

perhaps can simply be replicated by another subnet, or the same subnet taken over by someone else?

• On chain revenue sharing

Subnets generating revenue from their products should have a percentage AUTOMATICALLY flowing back to alpha holders and TAO stakers

Rayon Labs already does something like this with their auto-staking buyback on SN64 (Chutes) with their inference revenue; It should be standard not optional

• Headless subnet architecture. Remove the dependency on a single founding team entirely

The subnet operates as a protocol-level commodity where anyone can contribute compute and development. No one person or team can pull the plug

• Vesting schedules on founder emission allocation

Founders shouldn't be able to sell emissions as they receive them. Vest over 12-24 months minimum so there's a structural commitment to building long term rather than farming and leaving

This is somewhat entrenched with dTAO but there are loopholes to this, especially in a bull market

• Subnet governance minimums

Require subnets above a certain market cap or emission share to have multi-sig control, public roadmaps, and regular reporting to stakers

Treat it like a public company once you're managing other people's capital

• Exit penalties or cooldown periods

If a founding team wants to exit they face a cooldown where their locked tokens are gradually released rather than dumped. This gives stakers time to reprice and exit before the team does

• Insurance or protection pools funded by a small percentage of subnet emissions that pay out to alpha holders if a founding team exits or the subnet goes dark. Spreads the risk across the network

• Portable subnet infrastructure

Build the protocol so that if a team leaves the compute layer, model weights, and validation logic remain functional and can be picked up by another team or run autonomously. The subnet survives the founder

• Reputation and track record scoring on chain

Founders who have successfully run subnets long term and honoured their commitments get visible credibility scores. New subnet launches from unproven teams get flagged so stakers know the risk profile before they buy in.

None of these are silver bullets on their own but stacking several of them together would go a long way toward making subnet alpha tokens actually investable and keeping value accruing to TAO rather than leaking out every time a team decides to leave

I've seen my fair share of fraud in listed organisations as an auditor to know the problem comes from human error, be it "intentional" or otherwise

Code is law; "trust me bro" back of the napkin handshakes won't cut it anymore

Follow @TAOInstitute_ - the framework we are building will contribute to the gaps identified